June 2001 - Wine-Devel - list.winehq.org

Re: Re: virtual memory problems with Linux 2.4.5
by Mike McCormack June 11, 2001

June 11, 2001

i'm using Suse linux 7.0 with the following files: /usr/lib/libg++-3-libc6.1-2-2.8.1.3.so /lib/libc.so.6 i've upgrade the kernel and various bits and pieces, but not the C libraries. i think there was a discussion about changes to the mmap interface on KT at some time... Mike Original message from: Alexandre Julliard <julliard(a)winehq.com> > >Mike McCormack <mike_mccormack(a)start.com.au> writes: > >> i have found a problem with Wine and the Linux 2.4.5 Kernel. >> The Kernel no longer return EINVAL when offsets are not page >> aligned... it simply does the mapping ignoring the lower bits of the >> offset. This small patch fixes that problem. > >That's a glibc bug IMO. What version are you using? > >-- >Alexandre Julliard >julliard(a)winehq.com >. > ------------------------------------------ mailto:Mike_McCormack(a)looksmart.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

1 0

Re: winapi_extract and winapi_fixup
by Alexandre Julliard June 11, 2001

June 11, 2001

Patrik Stridvall <ps(a)leissner.se> writes: > Here is some new tools that is and will be > more useful in the future. If anybody > wishes to test them they should consider > them prealpha. I'd prefer if you put everything inside the winapi_check directory. Having 4 different directories for such similar tools is a bit overkill IMO. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: FYI: ControlSpy has stopped working
by Guy L. Albertelli June 11, 2001

June 11, 2001

-----Original Message----- From: Alexandre Julliard <julliard(a)winehq.com> To: andi(a)rhlx01.fht-esslingen.de <andi(a)rhlx01.fht-esslingen.de> Cc: Gerard Patel <gerard.patel(a)nerim.net>; Wine Devel <wine-devel(a)winehq.com> Date: Monday, June 11, 2001 4:52 PM Subject: Re: FYI: ControlSpy has stopped working >- ControlSpy refresh (?? needs more details) > The ControlSpy problem seems to be related to not repainting children windows of the rebar control. The attached traces show (+x11drv,+win,+rebar) for the 6/1 and 6/9 CVS. The 6/9 has submitted patch for zero client rect applied. If you want the full traces, let me know. Thanks, Guy Albertelli

1 0

Problems with GetFileVersionInfo[Size]()
by Malte.Starostik＠t-online.de June 11, 2001

June 11, 2001

Hi, The Shockwave ActiveX control I'm loading tries to read the VERSIONINFO resource from my winelib app. AFAICS it subsequently crashes because that fails. Anyway, there seems to be a problem: char path[PATH_MAX]; GetModuleFileName(inst, path, sizeof(path)); kdDebug() << "GetModuleFileName: " << path << endl; DWORD dummy; kdDebug() << "GetFileVersionInfoSize: " << GetFileVersionInfoSize(path, &dummy) << endl; kdDebug() << "GetLastError: " << GetLastError() << endl; outputs activexproxy: GetModuleFileName: Y:\src\kde\build\kdenonbeta\activexproxy\proxy\activexproxy activexproxy: GetFileVersionInfoSize: 0 activexproxy: GetLastError: 0 which means it obviously doesn't find the resource. (btw, why is GetLastError() 0 here?) However, this code suceeds (inst is the HINSTANCE passed to WinMain()): HRSRC ver = FindResource(inst, MAKEINTRESOURCE(VS_VERSION_INFO), MAKEINTRESOURCE(VS_FILE_INFO)); kdDebug() << "FindResource: " << ver << endl; int size = SizeofResource(inst, ver); kdDebug() << "SizeofResource:" << size << endl; HGLOBAL mem = LoadResource(inst, ver); void *buf, *block = LockResource(mem); UINT len; VerQueryValue(block, "StringFileInfo\\040904b0\\FileDescription", &buf, &len); kdDebug() << "Description: " << reinterpret_cast<WCHAR *>(buf) << endl; FreeResource(mem); the output is: activexproxy: FindResource: 1078760168 activexproxy: SizeofResource:608 activexproxy: Description: ActiveX-Proxy for Konqueror So basically the code GetFileInfoVersion[Size]() use to load a library fail if the given path is the winelib app (the symlink to wine) itself :( Thanks, -Malte -- Wars don't determine who's right, only who's left.

1 0

New Application Database is ONLINE
by Jeremy Newman June 11, 2001

June 11, 2001

Hey all! The 'shiny' new Wine Application Database is up and ready for your comments. We will be spending the rest of the day getting the initial category structure and 1st round off applications in. Once we have enough input, and we fell stable enough, I will link the AppDB to Winehq.com and post to wine-users. http://wine.codeweavers.com/appdb/ There is one sample app in so far so you can see what we mean. in: Main > Utilities > Compression > WinZip Use that as a guideline for how this is going to work. Go at it guys! -- _ _WebGeek/NetAdmin CodeWeavers -= http://www.codeweavers.com | \| |_____ __ ___ __ __ _ _ _ -= http://jnewman.codeweavers.com | .` / -_) V V / ' \/ _` | ' \ -= mailto:jnewman(a)codeweavers.com |_|\_\___|\_/\_/|_|_|_\__,_|_||_| -= ICQ: 1842980 Yahoo: laxdragon

1 0

New Application Database is ONLINE
by Jeremy Newman June 11, 2001

June 11, 2001

Hey all! The 'shiny' new Wine Application Database is up and ready for your comments. We will be spending the rest of the day getting the initial category structure and 1st round off applications in. Once we have enough input, and we fell stable enough, I will link the AppDB to Winehq.com and post to wine-users. http://wine.codeweavers.com/appdb/ There is one sample app in so far so you can see what we mean. in: Main > Utilities > Compression > WinZip Use that as a guideline for how this is going to work. Go at it guys! -- _ _WebGeek/NetAdmin CodeWeavers -= http://www.codeweavers.com | \| |_____ __ ___ __ __ _ _ _ -= http://jnewman.codeweavers.com | .` / -_) V V / ' \/ _` | ' \ -= mailto:jnewman(a)codeweavers.com |_|\_\___|\_/\_/|_|_|_\__,_|_||_| -= ICQ: 1842980 Yahoo: laxdragon

1 0

wine binary loader problem - incorrect code at entry point
by Mike McCormack June 11, 2001

June 11, 2001

Hi, Using the latest CVS, i'm trying to run the Adobe acrobat installer program (ar500enu.exe) and it crashes without doing anything at all. i have traced the problem to bad code at the program's entry point... which presumably means that the wrong piece of the program was mmapped into memory. i have run this program before using wine, (not recently though) and it has worked. Other programs (such as the netscape installer) at least start up and get going before crashing in a similar way. Mike from the debugger: 0x0040c110 (<Debugged process>.EntryPoint in <Debugged process>): decl 0x0(%edi) Using Sang Cho's disassembler i get: //******************** Program Entry Point ******** :0040C110 55 push ebp Disassembly of File: ./c/ar500enu.exe T.DateStamp = 351A66B8: Thu Mar 26 06:31:20 1998 Code Offset = 00000400, Code Size = 00010E00 Data Offset = 00011800, Data Size = 00003200 Number of Objects = 0005 (dec), Imagebase = 00400000h Object01: .text RVA: 00001000 Offset: 00000400 Size: 00010E00 Flags: 60000020 Object02: .rdata RVA: 00012000 Offset: 00011200 Size: 00000600 Flags: 40000040 Object03: .data RVA: 00013000 Offset: 00011800 Size: 00003200 Flags: C0000040 Object04: .idata RVA: 0001A000 Offset: 00014A00 Size: 00001000 Flags: C0000040 Object05: .rsrc RVA: 0001B000 Offset: 00015A00 Size: 0000A400 Flags: 40000040 and from the trace i get: trace:module:PE_LoadImage loading C:\ar500enu.exe trace:module:map_image mapped PE file at 0x400000-0x426000 trace:module:map_image mapping section .text at 0x401000 off 400 size 10e00 flags 60000020 trace:module:map_image mapping section .rdata at 0x412000 off 11200 size 600 flags 40000040 trace:module:map_image mapping section .data at 0x413000 off 11800 size 3200 flags c0000040 trace:module:map_image clearing 0x416200 - 0x417000 trace:module:map_image mapping section .idata at 0x41a000 off 14a00 size 1000 flags c0000040 trace:module:map_image mapping section .rsrc at 0x41b000 off 15a00 size a400 flags 40000040 trace:module:load_library loaded ntdll.dll 0x403512c4 40050000 ------------------------------------------ mailto:Mike_McCormack(a)looksmart.com.au ph +82 16 430 0425 __________________________________________________________________ Get your free Australian email account at http://www.Looksmart.com.au

1 0

[Problem report] X11DRV_ScrollWindowEx
by Gerard Patel June 10, 2001

June 10, 2001

This function hangs with the following apps (at least) - Notepad Win 3.11 - Notepad NT 4 - Eudora 1.5 In all cases, the problem appears when a text has several pages and I try to scroll toward the beginning of text. The problem disappears if I comment out the calls to X11DRV_StartGraphicsExposure and X11DRV_EndGraphicsExposure. The apps seem to work just as well. The hanging occurs because there is no X event generated in X11DRV_EndGraphicsExposure. I don't quite see why there should be one in this case. Gerard

1 0

FYI: ControlSpy has stopped working
by Guy L. Albertelli June 10, 2001

June 10, 2001

Using the current CVS (6/9), certain ControlSpy modules (REBAR and COMBOBOXEX) no longer seem to work, even with the native versions of COMCTL32. The internal child windows in the left pane are no longer drawn. Seems to have occured sometime since 6/1. I will look into it tomorrow after some rest. Strange. Guy Albertelli <<galberte(a)neo.lrun.com>>

2 1

Re: FYI: ControlSpy has stopped working
by Gerard Patel June 10, 2001

June 10, 2001

At 11:56 PM 09/06/2001 -0400, you wrote: >Using the current CVS (6/9), certain ControlSpy modules (REBAR and >COMBOBOXEX) no longer seem to work, even with the native versions of >COMCTL32. The internal child windows in the left pane are no longer drawn. >Seems to have occured sometime since 6/1. I will look into it tomorrow after >some rest. Strange. Strange ??? After this patch : ChangeSet ID: 991691718458116964902729 CVSROOT: /opt/cvs-commit Module name: wine Changes by: julliard(a)wine2. 01/06/04 17:55:18 (...) Log message: Create an X window for every window, including children. I see a lot of problems in most of my test apps after the implementation of this evil scheme. I have already sent several patches to work around some of these problems, two have been applied, one has been ignored so far, but even with this last Wine's status is beyond the usual breakage level. Gerard

2 1

Help needed
by Renukaprasad, MR June 9, 2001

June 9, 2001

Hi, We are planning to port wine onto Tru64 Unix. I have some basic Questions. Can you tell me how to go about it. How do we start. How big is the task and how to start about it.. Thanks And Regards, Renuka

1 0

Re: VirtualProtect() fixup
by Alexandre Julliard June 8, 2001

June 8, 2001

520053692817-0001(a)t-online.de (Andreas Mohr) writes: > (you do know that I disagree on these things, right ? after all Wine emulates > *Windows*, and if Windows does it that way, then we should do it, too, for > heaven's sake, especially if it's not really painful) There is no such thing as "Windows", there are multiple versions that do different things, and we have to choose what we emulate. In this case it's clear that it makes more sense to emulate the NT behavior, since this is what is documented. Of course we could do a winver check, but if we do this for every single difference between Win98 and NT, we'll end up with a mess of spaghetti code with more version checks than actual code. So winver checks must be limited to cases where they are demonstrably necessary for a real-world application. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: VirtualProtect() fixup
by Alexandre Julliard June 8, 2001

June 8, 2001

520053692817-0001(a)t-online.de (Andreas Mohr) writes: > What isn't the case ? Both fixups ? Or only one of these ? The return value: NT returns a BOOL, 0 or 1. I haven't checked the old protection behavior; but anyway this doesn't belong in VirtualProtect, if we have to ignore the execute permission it must be done consistently throughout the code. > Well, to be exact, no application. > I just found this out by accident. That's what I suspected ;-) > IMHO the probability that some applications exist that are too lazy > to supply their own old_prot variable and that check the return value instead > is rather high, so I think that patch is necessary. We can't start breaking things based only on your suspicion that someone somewhere will have used a feature that isn't documented at all (plus there is a documented way of doing exactly the same thing). If you have an app that needs it sure, otherwise we'll just keep returning a BOOL, as the function is documented to do. > Maybe we should add a winver check, though. No thanks. -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: wcmd fix
by Francois Gouget June 8, 2001

June 8, 2001

On Fri, 8 Jun 2001, Andreas Mohr wrote: > Hi all, > > file_size.XXXPart should be file_size.s.XXXPart. > > Thanks to Lasse Melsaeter for spotting this. That's the 'Dummy struct name problem in wcmd' I sent this morning. But actually whether there should an 's.' depends on the compiler (for Winelib applications) so you should test for NONAMELESSSTRUCT as I did in my patch. -- Francois Gouget fgouget(a)free.fr http://fgouget.free.fr/ If it stinks, it's chemistry. If it moves, it's biology. If it does not work, It's computer science.

1 0

Re: VirtualProtect() fixup
by Alexandre Julliard June 8, 2001

June 8, 2001

Andreas Mohr <a.mohr(a)mailto.de> writes: > VirtualProtect had two big problems: > > - undocumented return value > - should return protection flags only instead of whole flag setting This isn't the case under NT. Which application depends on that? -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: Fwd: Re: WINE fix for HOMM3 [bluefire@zauberwald.net]
by Uwe Bonnes June 7, 2001

June 7, 2001

>>>>> "BlueFire" == BlueFire <bluefire(a)zauberwald.net> writes: BlueFire> Now with the missing diff..... sorry... On Thu, 07 Jun 2001 BlueFire> 11:37:14 BlueFire wrote: >> This is messed up (totally unreadable) formatting. Anyway, how about >> doing a diff -u of your change, and then submitting that patch (in >> more readable form) either to Debian's BTS, or directly to >> wine-patches(a)winehq.com? BlueFire> Sorry, I'm not familiar with either diff or the debian BlueFire> bts. That's why I tried mailing the stuff to you (and to BlueFire> wine-patches). I did a BlueFire> diff /tmp/dos_fs.c dos_fs.c > dos_fs.c.diff Hallo, find some help at http://www.winehq.com/dev.shtml BlueFire> diff -u is preferred due to it's better readability. Appending the plain file makes it easier for others to read your patch. But don't " cut and paste", as that will ofter reformat and make the patch a pain to apply. Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

1 0

Question about the new window design
by Gerard Patel June 6, 2001

June 6, 2001

Since the patch of 4 June, all windows are created as X windows. This makes one of my test apps to crash; it's creating a child window in the WM_NCCREATE event handling of a window. When the child window is created, X bombs because the X window parent does not exist yet - Wine is trying to call XCreateWindow with parent = 0. What is the best way of fixing this ? Gerard

1 0

Re: stack limit
by Alexandre Julliard June 6, 2001

June 6, 2001

Stephen Crowley <stephenc(a)placemark.com> writes: > Works like a charm! Thanks. I'm curious though.. I set the stack size > to 64 megs and the application immediately uses up that much ram. Is > that the default behaviour on windows? What stops it from behaving > like a normal stack and just allocating what's needed from the heap? The ram isn't actually used, it's only reserved from the address space with an anonymous mapping. Only the pages that are really accessed will use physical memory/swap, so it doesn't use any more ram than a normal stack. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: stack limit
by Alexandre Julliard June 6, 2001

June 6, 2001

Stephen Crowley <stephenc(a)placemark.com> writes: > On Mon, Jun 04, 2001 at 12:02:09PM -0700, Alexandre Julliard wrote: > > There is no way to do this at the moment; but it shouldn't be too hard > > to add support for a 'stack' statement in the application spec file. > > Thanks. Where would I look to change this in the source? I'm using some linear > regression DLLs under linux and we have to deal with some fairly large > matrices. Try this: Index: scheduler/process.c =================================================================== RCS file: /opt/cvs-commit/wine/scheduler/process.c,v retrieving revision 1.154 diff -u -r1.154 process.c --- scheduler/process.c 2001/06/06 20:24:13 1.154 +++ scheduler/process.c 2001/06/06 22:46:20 @@ -319,12 +319,6 @@ LPTHREAD_START_ROUTINE entry; WINE_MODREF *wm; - /* build command line */ - if (!ENV_BuildCommandLine( main_exe_argv )) goto error; - - /* create 32-bit module for main exe */ - if (!(current_process.module = BUILTIN32_LoadExeModule( current_process.module ))) goto error; - /* use original argv[0] as name for the main module */ if (!main_exe_name[0]) { @@ -498,7 +492,6 @@ { if (PE_HEADER(current_process.module)->FileHeader.Characteristics & IMAGE_FILE_DLL) ExitProcess( ERROR_BAD_EXE_FORMAT ); - stack_size = PE_HEADER(current_process.module)->OptionalHeader.SizeOfStackReserve; goto found; } @@ -512,6 +505,13 @@ _EnterWin16Lock(); found: + /* build command line */ + if (!ENV_BuildCommandLine( main_exe_argv )) goto error; + + /* create 32-bit module for main exe */ + if (!(current_process.module = BUILTIN32_LoadExeModule( current_process.module ))) goto error; + stack_size = PE_HEADER(current_process.module)->OptionalHeader.SizeOfStackReserve; + /* allocate main thread stack */ if (!THREAD_InitStack( NtCurrentTeb(), stack_size )) goto error; Index: tools/winebuild/README =================================================================== RCS file: /opt/cvs-commit/wine/tools/winebuild/README,v retrieving revision 1.4 diff -u -r1.4 README --- tools/winebuild/README 2000/12/26 01:22:34 1.4 +++ tools/winebuild/README 2001/06/06 22:46:21 @@ -6,6 +6,7 @@ [file WINFILENAME] [mode dll|cuiexe|guiexe|cuiexe_unicode|guiexe_unicode] [heap SIZE] +[stack SIZE] [init FUNCTION] [import [IMP_FLAGS] DLL] [rsrc RESFILE] @@ -39,6 +40,9 @@ "heap" is the size of the module local heap (only valid for Win16 modules); default is no local heap. + +"stack" is the stack size for Win32 exe modules, in kilobytes; default +size is 1024 (1Mb stack). "file" gives the name of the Windows file that is replaced by the builtin. <name>.DLL is assumed if none is given. (This is important Index: tools/winebuild/build.h =================================================================== RCS file: /opt/cvs-commit/wine/tools/winebuild/build.h,v retrieving revision 1.17 diff -u -r1.17 build.h --- tools/winebuild/build.h 2001/05/22 19:55:51 1.17 +++ tools/winebuild/build.h 2001/06/06 22:46:21 @@ -161,6 +161,7 @@ extern int DLLHeapSize; extern int UsePIC; extern int debugging; +extern int stack_size; extern int nb_debug_channels; extern int nb_lib_paths; Index: tools/winebuild/main.c =================================================================== RCS file: /opt/cvs-commit/wine/tools/winebuild/main.c,v retrieving revision 1.17 diff -u -r1.17 main.c --- tools/winebuild/main.c 2001/05/22 19:55:51 1.17 +++ tools/winebuild/main.c 2001/06/06 22:46:21 @@ -27,6 +27,7 @@ int Limit = 0; int DLLHeapSize = 0; int UsePIC = 0; +int stack_size = 0; int nb_entry_points = 0; int nb_names = 0; int nb_debug_channels = 0; Index: tools/winebuild/parser.c =================================================================== RCS file: /opt/cvs-commit/wine/tools/winebuild/parser.c,v retrieving revision 1.18 diff -u -r1.18 parser.c --- tools/winebuild/parser.c 2001/05/22 19:55:51 1.18 +++ tools/winebuild/parser.c 2001/06/06 22:46:21 @@ -559,6 +559,12 @@ if (!IsNumberString(token)) fatal_error( "Expected number after heap\n" ); DLLHeapSize = atoi(token); } + else if (strcmp(token, "stack") == 0) + { + token = GetToken(0); + if (!IsNumberString(token)) fatal_error( "Expected number after stack\n" ); + stack_size = atoi(token); + } else if (strcmp(token, "init") == 0) { if (SpecType == SPEC_WIN16) Index: tools/winebuild/spec32.c =================================================================== RCS file: /opt/cvs-commit/wine/tools/winebuild/spec32.c,v retrieving revision 1.28 diff -u -r1.28 spec32.c --- tools/winebuild/spec32.c 2001/05/22 19:55:51 1.28 +++ tools/winebuild/spec32.c 2001/06/06 22:46:21 @@ -656,7 +656,10 @@ fprintf( outfile, " %ld,\n", page_size ); /* SizeOfHeaders */ fprintf( outfile, " 0,\n" ); /* CheckSum */ fprintf( outfile, " 0x%04x,\n", subsystem ); /* Subsystem */ - fprintf( outfile, " 0, 0, 0, 0, 0, 0,\n" ); + fprintf( outfile, " 0,\n" ); /* DllCharacteristics */ + fprintf( outfile, " %d, 0,\n", stack_size*1024 ); /* SizeOfStackReserve/Commit */ + fprintf( outfile, " %d, 0,\n", DLLHeapSize*1024 );/* SizeOfHeapReserve/Commit */ + fprintf( outfile, " 0,\n" ); /* LoaderFlags */ fprintf( outfile, " %d,\n", IMAGE_NUMBEROF_DIRECTORY_ENTRIES ); /* NumberOfRvaAndSizes */ fprintf( outfile, " {\n" ); fprintf( outfile, " { %s, %d },\n", /* IMAGE_DIRECTORY_ENTRY_EXPORT */ -- Alexandre Julliard julliard(a)winehq.com

2 1

Re: Question about the new window design
by Alexandre Julliard June 6, 2001

June 6, 2001

Gerard Patel <gerard.patel(a)nerim.net> writes: > This makes one of my test apps to crash; it's creating a child window > in the WM_NCCREATE event handling of a window. When the > child window is created, X bombs because the X window parent > does not exist yet - Wine is trying to call XCreateWindow with > parent = 0. > > What is the best way of fixing this ? The X window should be created before the WM_NCCREATE. I'm planning to fix it, but it needs to correctly handle style changes happening in WM_NCCREATE (otherwise we get the managed tooltips bug), and this needs a bit more work. -- Alexandre Julliard julliard(a)winehq.com

1 0

Re: uname(1) on UnixWare (Solaris)
by Alexandre Julliard June 6, 2001

June 6, 2001

Bang Jun-Young <bjy(a)mogua.org> writes: > I've nearly completed new DLL configuration code in configure.in. > Before making it available, I need to know what uname(1) command > prints out on UnixWare (and Solaris?). Why do you need that? Using uname(1) in a configure script is almost certainly the wrong thing to do. -- Alexandre Julliard julliard(a)winehq.com

4 9

Re: [PATCH] Flag unicode structure for user create window
by Uwe Bonnes June 6, 2001

June 6, 2001

>>>>> "Gerard" == Gerard Patel <gerard.patel(a)nerim.net> writes: Gerard> In the CreateWindow user driver function, it is not enough to Gerard> test the window procedure to know if the cs structure is unicode Gerard> or not : the window can be subclassed, in the cbt hook, for Gerard> example. Gerard> ChangeLog: Hallo, yesterday I tried to find why the findfile95 dialog only displayed the first character ( "o" versus "open"). Your patch fixed it :-) Bye -- Uwe Bonnes bon(a)elektron.ikp.physik.tu-darmstadt.de Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt --------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------

1 0

RE: uname(1) on UnixWare (Solaris)
by Patrik Stridvall June 6, 2001

June 6, 2001

> On Tue, Jun 05, 2001 at 12:18:00AM -0700, Francois Gouget wrote: > > On Tue, 5 Jun 2001, Bang Jun-Young wrote: > > > 1. Configure message is wrong: > > > checking whether we can build a Linux dll... yes > > > I see this even on FreeBSD and NetBSD. :( It's not a > Linux dll, > > > but an FreeBSD/NetBSD dll. > > > > Maybe this message should be changed, yes. But I don't think you > > should put a `uname` there. > > Could you explain me a bit further why we shouldn't do that? As far as the question above is concerned see Francois's reply. I fully agree with him. However regardless of this, uname shouldn't be used (at least not directly). Autoconf provides a standard way to do this (which BTW happends to use uname). It can be used as below. AC_CANONICAL_HOST case $host_os in *cygwin* ) CYGWIN=yes;; * ) CYGWIN=no;; esac It requires two files config.sub and config.guess be distribute along with Wine but that is not really a problem. Perhaps we need to do this anyway. Autoconf 2.50 has discontinued AC_CYGWIN so CYGWIN support seems to require it regardless. Actually the code above is how this is done with Autoconf 2.50. However making CYGWIN an exception is one thing, doing it in other cases is quite another thing. IMHO you will need to provide a _really_ convincing argument that we should make another exception assuming of course we make the first one for CYGWIN which we haven't done yet either.

2 1

RE: uname(1) on UnixWare (Solaris)
by Patrik Stridvall June 6, 2001

June 6, 2001

> On Tue, Jun 05, 2001 at 11:51:19PM -0700, Francois Gouget wrote: > > Maybe it was not your intention but I was afraid that we > would end up > > with something like: > > "checking whether we can build a `uname` dll..." > > and then a few lines down: > > "checking whether we can build a `uname` dll..." > > > > The problem is then you don't know exactly where the > message comes > > from. > > Sorry but I don't catch what you mean quite well. It's obvious which > message is printed on the screen, isn't it? Yes, but this means that the text no longer describes what it is trying to check. If somebody runs it on a OS called Foo it might trick the him/her into thinking that we know about and have added special handling for OS Foo, which we haven't. > What do you think about replacing "Linux dll" with "GNU style > ELF dll"? Acceptable, but it is not my decision.

1 0

Is this a bug?
by Ian Pilcher June 6, 2001

June 6, 2001

While poking around in gdi.h, I came upon the following: static inline INT WINE_UNUSED INTERNAL_XWSTODS(DC *dc, INT width) { FLOAT floatWidth; /* Perform operation with floating point */ floatWidth = (FLOAT)width; INTERNAL_XWSTODS_FLOAT(dc, &floatWidth); /* Round to integers */ return GDI_ROUND(floatWidth); } ... and a little later ... static inline INT WINE_UNUSED INTERNAL_YWSTODS(DC *dc, INT height) { FLOAT floatHeight; /* Perform operation with floating point */ floatHeight = (FLOAT)height; INTERNAL_XWSTODS_FLOAT(dc, &floatHeight); /* Round to integers */ return GDI_ROUND(floatHeight); } Shouldn't one of these (probably INTERNAL_YWSTODS) be calling INTERNAL_YWSTODS_FLOAT instead of INTERNAL_XWSTODS_FLOAT? -- ======================================================================== Ian Pilcher ian.pilcher(a)home.com ========================================================================

2 1